Bogumił Jakubiak

IMPLEMENTATION OF DATA MINING TECHNIQUES FOR METEOROLOGICAL APPLICATIONS

The CrossGrid project is one of the ongoing research projects involving GRID technology. One of the main tasks in the Meteorological applications package is the implementation of data mining systems for the analysis of operational and reanalysis databases of atmospheric circulation patterns. Previous parallel data mining algorithms reported in the literature focus on parallel computers with predetermined resources (processing units) and high-performance communications. The main goal in this project is designing adaptive schemes for distributing data and computational load according to the changing resources available for each GRID job submitted. In this paper, some preliminary work regarding two different data mining algorithms (self organizing maps and smoothing filters) is presented. These techniques can be used in combination with databases of observati ons to provide downscaled local forecasts from operative model outputs. This is a more general framework to look at data mining techniques from the meteorological point of view.

The Effect of the Physical Parameterizations and the Land Surface on Rainfall in Poland

AbstractHigh-resolution numerical experiments were conducted over two separate months to study the effect of different physical parameterizations and different representations of the land surface on the prediction of rainfall events in Poland. The Coupled Ocean–Atmosphere Mesoscale Prediction System (COAMPS) was used with 2-km grid spacing. Four sets of forecast experiments were performed. The control experiment used a slab model for the surface energy budget, a Lin-based moist physics parameterization, and a Mellor and Yamada (MY) based turbulent kinetic energy (TKE) parameterization; the Louis experiment used a version of the Louis TKE parameterization in place of MY; the LSM experiment used the Noah land surface model (LSM) in place of the slab model; and the Thompson experiment used the Thompson microphysics parameterization instead of the Lin-based microphysics. The forecasts were validated against surface and upper-air observations, as well as radar reflectivity. The Louis parameterization yielded i...

Impact of Improved Land-Surface Model Coupled to NWP System on Convective Boundary Layer Variables

Land surface parameterizations in numerical weather prediction models direct the exchange of energy between the land surface and the atmosphere. The main purpose of this paper is to gain insight into how surface heterogeneity can influence the modeled convective boundary layer, and whether the complexity of the modeled land-atmosphere interactions can improve forecasts of convective precipitation. To achieve this the NOAH land-surface model has been coupled to the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS). Selected cases of fair weather convective conditions and convectivly produced rain episodes were used to evaluate the behavior of a land-surface model coupled to a high-resolution numerical weather prediction system. Our experiments are designed such that different surface conditions (wet soil during rain episodes and dry soil conditions during fair weather days) and different atmospheric situations could be used to further our understanding of the behavior of the model by comparing model solutions to surface observations.

Mesoscale Convective Systems Monitoring on the Basis of MSG Data – Case Studies

Abstract Analysis described in the paper were made in the frame of the PROZA (Operational decisionmaking based on atmospheric conditions, http://projekt-proza.pl/) project co-financed by the European Union through the European Regional Development Fund. One of its tasks was to develop an operational forecast system, which is going to support different economies branches like forestry or fruit farming by reducing the risk of economic decisions with taking into consideration weather conditions. The main purpose of the paper is to describe the method of the MCSs (Mesoscale Convective Systems) tracking on the basis of the MSG (Meteosat Second Generation) data. Until now several tests were performed. The Meteosat satellite images in selected spectral channels collected for Central Europe Region for May 2010 were used to detect and track cloud systems recognized as MCSs in Poland. The ISIS tracking method was applied here. First the cloud objects are defined using the temperature threshold and next the selected cells are tracked using principle of overlapping position on consecutive images. The main benefit of using a temperature threshold to define cells is its efficiency. During the tracking process the algorithm links the cells of the image at time t to the one of the following image at time t+dt that correspond to the same cloud system. Selected cases present phenomena, which appeared at the territory of Poland. They were compared to the weather radar data and UKMO UM (United Kingdom MetOffice Unified Model) forecasts. The paper presents analysis of exemplary MCSs in the context of near realtime prediction system development and proves that developed tool can be helpful in MCSs monitoring.

Processing of 3D Weather Radar Data with Application for Assimilation in the NWP Model

Abstract The paper is focused on the processing of 3D weather radar data to minimize the impact of a number of errors from different sources, both meteorological and non-meteorological. The data is also quantitatively characterized in terms of its quality. A set of dedicated algorithms based on analysis of the reflectivity field pattern is described. All the developed algorithms were tested on data from the Polish radar network POLRAD. Quality control plays a key role in avoiding the introduction of incorrect information into applications using radar data. One of the quality control methods is radar data assimilation in numerical weather prediction models to estimate initial conditions of the atmosphere. The study shows an experiment with quality controlled radar data assimilation in the COAMPS model using the ensemble Kalman filter technique. The analysis proved the potential of radar data for such applications; however, further investigations will be indispensable.